Configuring the Intelligent Information Matching Flow

The following figure shows the preconfigured intelligent information matching flow.

This flow implements the following functions:

1. The response diagram element plays the greeting to the customer. Customize the voice files.
2. The Interface Invoking diagram element at position 1 separates the texts sent by the client by spaces based on the interaction records, and records whether the texts are sent by the customer or agent.
3. The Semantic Recognition diagram element identifies the interaction information as follows:
   • If sensitive words exist, the sensitive word prompt is triggered. The Interface Invoking diagram element at point 3 assigns the sensitive words to flow variables.
   • If no sensitive word exists and the intention is matched, the Interface Invoking diagram element at point 4 is entered, and the matched intention name is used to invoke the Q&A interface to obtain a recommended reply.
   • If an error such as timeout or unknown intention occurs, the Interface Invoking diagram element at point 2 is invoked. The interaction information is directly used to invoke the Q&A interface to obtain a recommended reply.
4. The response diagram element returns the recommended replies and sensitive words.
5. After a reply is triggered, the sensitive words are empty and the next cycle starts.

Modify the greeting voice templates and invoke the specific recommendation and knowledge obtaining interfaces in an Interface Invoking diagram element to implement the information matching function.

• Greeting Resource Templates
  Resource templates in the first response of the information matching flow can be configured and modified by the customer or preconfigured by the business personnel based on the customer requirements.
• Interface Invoking Diagram Element
  The information matching flow supports user-defined modification and update of invoked diagram element types and settings based on business requirements.
• Response Templates Returned to the Semantic Recognition Invoker
  Response templates can be configured by the customer or configured by the business personnel based on the customer requirements. After a template is configured, the backend automatically reads the parameter information.